Explainable AI-Enabled IoT System for Automated Crop Health Monitoring and Precision Agriculture

The present trend in precise farming relies on AI and IoT to improve our sensoring abilities and optimize farm operations. However, traditional AI algorithms can be difficult to believe and adopt since they tend to seem unmeaningful and nonsensical to humans. In this paper, we introduce an Explainable AI-assisted IoT edge solution for monitoring and optimizing our farm and crop activities through precise farming. This proposed use of AI combines distributed sensors with data from IoT on soil moisture content, temperature, humidity, and nutrient concentration with an XGBoost predictive model running on an IoT edge to provide real-time judgments for crop health. SHAP (SHapley Additive explanations) explains this predictive modeling process by providing signaling values based on environmental factors with the largest contributions to health outcomes. The experimental outcomes show that our proposed IoT system provides maximum predictive classification with 94.5 overall accuracy compared to existing CNN and RNN mechanism techniques and is suitable for low-latency real-time inference tasks. It is further clarified that soil moisture and temperature are major predictors to differentiate healthy from stressed and sick crops. Our proposed IoT mechanism based on deep learning provides actionable intelligence to farms to effectively act at the right time for healthy and sustainable farming. The framework lays a base for future multi-crop and multi-location use in smart agriculture.